On a class of designs obtained by composition of ovaloids (Q2563493): Difference between revisions

Let \(O\) be an ovoid of the projective 3-space \(\text{PG} (3,q^2)\). Every three distinct points determine a unique plane section \(S\), which has \(q^2+1\) points in common with \(O\). We can, in an arbitrary way, consider these \(q^2+1\) points as the points of an ovoid \(O_S\) of \(\text{PG}(3,q)\), in which every three points again determine a unique plane section of \(q+1\) points. This way, we have a \(3-(q^4+1,q+1,1)\)-design. The authors now consider the special case that all \(O_S\) are isomorphic (in this case, the construction above is called the composition of \(O\) and \(O_S)\), and show that in general (i.e., up to a finite number of small cases), these designs do not admit subspaces (a subspace is a subset \(T\) of the set of points of \(O\) such that all \(q+1\) points of the unique block through any three points of \(T\) are contained in \(T)\) other than the subdesigns \(O_S\), except possibly if the cardinality of the subspace is \(q^2+1\) and \(O\) is not a quadric. It seems to me that the proof does not use the fact that all \(O_S\) are isomorphic and that the authors thus proved a stronger result. Finally, the possible exceptions are not classified.